ACE-Step

Docker app from SpaceInvaderOne's Repository

Overview

ACE-Step 1.5 - AI Music Generation. Generate full songs with vocals, instrumentals, and lyrics using a Diffusion Transformer. Supports text-to-music, remixing, cover generation, and LoRA fine-tuning. Requires NVIDIA GPU with CUDA support.

FIRST RUN: Models (~10GB) will be downloaded automatically on first start. This may take several minutes depending on your internet speed. Subsequent starts are instant.

SETTINGS GUIDE:

DiT Model - The core music generation model.

turbo (default): Fast generation in 8 steps. Best for most users.
turbo-rl: Turbo with reinforcement learning refinement.
sft: Higher quality, 50 steps (slower).
base: 50 steps with all features (extract, lego, complete).

Language Model - Controls lyrics understanding and chain-of-thought reasoning.

1.7B (default): Best balance of quality and VRAM. Recommended for 12-16GB GPUs.
0.6B: For GPUs with less than 12GB VRAM.
4B: Highest quality lyrics understanding. Requires 24GB+ VRAM.

Enable LLM - Whether to load the language model.

auto (default): Detects based on your GPU VRAM.
false: DiT-only mode. Faster startup, uses less VRAM, but disables thinking/sample features.
true: Force enable.

LM Backend - Engine for the language model.

pt (default): PyTorch native. Works on all GPUs including RTX 50-series.
vllm: Faster inference but may crash on RTX 50-series (Blackwell) GPUs.

CPU Offloading - Moves models between GPU and CPU to save VRAM.

auto (default): Offloads if GPU has less than 20GB VRAM.
false: Keep all models on GPU. Faster generation but uses ~12GB VRAM at idle.
true: Always offload. Slower but frees VRAM for other containers.

UI Language - Web interface language: English, Chinese, or Japanese.

Requirements

IMPORTANT: This image requires at least 20GB of free space in your Docker vDisk. Check Settings > Docker > Docker vDisk Size and increase if needed. Requires NVIDIA GPU with 8GB+ VRAM (12GB+ recommended for full features). Models (~10GB) are downloaded on first run to the mapped checkpoints volume.

Runtime arguments

Web UI: http://[IP]:[PORT:7860]/
Network: bridge
Shell: bash
Privileged: false
Extra Params: --gpus all --user root

Template configuration

Web UI PortPorttcp

Gradio Web UI and REST API port

Target: 7860
Default: 7860
Value: 7860

Model CheckpointsPathrw

AI model files (~10GB, auto-downloaded on first run)

Target: /app/checkpoints
Default: /mnt/user/appdata/ace-step/checkpoints
Value: /mnt/user/appdata/ace-step/checkpoints

Generated MusicPathrw

Generated music files output directory

Target: /app/gradio_outputs
Default: /mnt/user/appdata/ace-step/output
Value: /mnt/user/appdata/ace-step/output

DiT ModelVariable{3}

Diffusion model variant. Turbo=8 steps (fast), SFT=50 steps (quality), Base=50 steps (all features including extract/lego/complete).

Target: ACESTEP_CONFIG_PATH
Default: acestep-v15-turbo|acestep-v15-turbo-rl|acestep-v15-sft|acestep-v15-base
Value: acestep-v15-turbo

Language ModelVariable{3}

Chain-of-thought LM size. 1.7B recommended for 16GB VRAM. 4B needs 24GB+. 0.6B for low VRAM.

Target: ACESTEP_LM_MODEL_PATH
Default: acestep-5Hz-lm-1.7B|acestep-5Hz-lm-0.6B|acestep-5Hz-lm-4B
Value: acestep-5Hz-lm-1.7B

Enable LLMVariable{3}

Auto detects based on GPU VRAM. Set false for DiT-only mode (faster, no thinking/sample features).

Target: ACESTEP_INIT_LLM
Default: auto|true|false
Value: auto

LM BackendVariable{3}

pt (PyTorch native) is recommended for RTX 50-series. vllm (nano-vllm) is faster but may segfault on Blackwell GPUs.

Target: ACESTEP_LM_BACKEND
Default: pt|vllm
Value: pt

UI LanguageVariable{3}

Web interface language

Target: LANGUAGE
Default: en|zh|ja
Value: en

CPU OffloadingVariable{3}

auto = ACE-Step decides based on VRAM (offloads below 20GB). false = keep all models on GPU (faster, needs ~12GB idle VRAM). true = offload models to CPU between steps (slower, saves VRAM for shared GPU use).

Target: ACESTEP_OFFLOAD_CPU
Default: auto|false|true
Value: auto

Web UI Port (internal)Variable{3}

Internal port for Gradio server (should match the port mapping above)

Target: PORT
Default: 7860
Value: 7860

Batch SizeVariable{3}

Default generation batch size (1-8). Leave empty for auto (min(2, GPU max)).

Target: ACESTEP_BATCH_SIZE

CUDA Visible DevicesVariable{3}

Which GPU(s) to use. 0 = first GPU.

Target: CUDA_VISIBLE_DEVICES
Default: 0
Value: 0

Download SourceVariable{3}

Model download source. Auto tries HuggingFace first, falls back to ModelScope.

Target: ACESTEP_DOWNLOAD_SOURCE
Default: auto|huggingface|modelscope
Value: auto

Links

Template Docker Hub Project

Details

Repository

spaceinvaderone/ace-step:latest

Registry

https://hub.docker.com/r/spaceinvaderone/ace-step/

Last Updated2026-02-22

First Seen2026-02-22

Run ACE-Step on Unraid.

ACE-Step is listed in Community Apps for Unraid OS. Explore Unraid to build a flexible home server, NAS, or homelab.

Explore Unraid OS